HOXD-AS1 facilitates cell migration and invasion as an oncogenic lncRNA by competitively binding to miR-877-3p and upregulating FGF2 in human cervical cancer

CC tissues and the matched non-cancerous normal tissues were received from 40 patients in the Tianjin Baodi People’s Hospital from January 2018 to December 2019. The matched non-cancerous normal tissues were obtained at a distance of 5 cm from cancerous tissues. Clinical specimens were immediately frozen in liquid nitrogen and stored at − 80 °C for further experiments. The current study was carried out in accordance with the guidelines by the Ethics and Scientific Committee of Tianjin Baodi People’s Hospital. Written informed consent was obtained from all patients involved in this study.

Four CC cell lines (SiHa, CaSki, Hela, C-33A) and one normal cervical epithelial cell line (Ect1/E6E7) were purchased obtained from the Cell Bank of Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). Cells were cultured in 1640 or DMEM medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific), 100 U/ml penicillin, and 100 μg/ml streptomycin (Thermo Fisher Scientific). They were cultured at 37 °C in a humidified incubator with 5% CO₂.

HOXD-AS1 siRNAs, miR-877-3p mimics and inhibitor, negative control oligos were all ordered from GenePharma (Shanghai, China). SiRNAs, miRNA mimics or inhibitor were transfected into CC cells with Lipofectamine™ 2000 Transfection Reagent (Thermo Fisher Scientific) according to the manufacturer’s standard protocol.

Total RNAs from CC tissues or transfected cells were extracted by using Trizol reagent (Thermo Fisher Scientific). 2 μg of RNA were reverse transcribed to detect HOXD-AS1 and GAPDH levels using the general reverse transcription kit (Promega, Madison, WI, USA). 1 μg of RNA were reverse transcribed to detect examine miR-877-3p and U6B levels using the specific primer kit (MQPS0002242–1-100 and MQPS0000002–1-100, RiboBio, Guangzhou, China). qRT-PCR was performed on the Applied Biosystems 7900HT Fast Real-Time PCR System using a standard protocol. The results were normalized to GAPDH or U6B. All the experiments were performed in three times as per the manufacturer’s instructions. The expression fold changes were calculated using the 2^-ΔΔCt method. The primer sequences used for qRT-PCR were as follows:

FGF2:

5′-ATGTAGAAGATGTGACGCCG-3′

5′-GGTTCACGGATGGGTGTCT-3′

HOXD-AS1:

5′-TCCTCAGGTCTAAGGACGGG-3′

5′-AGCGGAAGAGTAGGTCTGGT-3′

GAPDH:

5′-GTCAAGGCTGAGAACGGGAA-3′

5′-AAATGAGCCCCAGCCTTCTC-3′

When the transfected cells were grown up to 90% confluence, cell monolayer was scraped in straight lines using an 1 ml pipette tip and washed with PBS twice. Then cells were further cultured in serum-free culture medium with 1% penicillin/streptomycin. Images were captured at 0 h or 48 h following the initial scratch at five random areas for each wound. ImageJ software (NIH, Bethesda, MD, USA) was used to calculate cell migration rate as: (the original wound width - the actual wound width) / (the original wound width).

Cell invasion ability was determined by transwell assay. Briefly, cells were seeded into 24-well upper chambers (Corning, NY, USA) pre-coated with Matrigel (Millipore, MA, USA) in serum-free medium. After 48 h, the invasive cells were fixed with 4% paraformaldehyde, followed by staining with 0.25% crystal violet solution (Sigma-Aldrich Co., St Louis, MO, USA) for 20 min. Then stained cells were observed and counted under an inverted microscope (Nikon, Japan), and five different microscopic views were selected randomly for analysis.

The wild type sequence of HOXD-AS1 was synthesized by BGI (The Beijing Genomics Institute, Beijing, China) and cloned into the pcDNA3.1 vector (Thermo Fisher Scientific). The wild type or mutant type of 3′-UTR region of FGF2 mRNA was cloned into the pmirGLO vector (Promega). The wild type or mutant truncated sequence of HOXD-AS1 was synthesized by BGI and cloned into the pmirGLO vector.

Hela cells were co-transfected with miR-877-3p mimics as well as wild type or mutant HOXD-AS1. Relative luciferase activity was measured on a dual-luciferase reporter system (Promega) as per the manufacturer’s instructions. Data were calculated as the ratio of renilla to firefly luciferase activity. Luciferase reporter assays to validate the direct binding of miR-877-3p to FGF2 3′-UTR were also performed as described above.

Total protein was extracted from cells using RIPA buffer (Beyotime Biotechnology, Nanjing, China) following the manufacturer’s instructions. Protein lysates were separated by 10% SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis) and transferred to PVDF membranes. The membranes were then blocked with PBS containing 0.1% Triton X-100 and 5% slim milk at 25 °C for 1 h, before being incubated with anti-FGF2 or anti-β-actin antibody (Santa Cruz, Dallas, TX, USA) at 4 °C overnight. After washing, the membranes were incubated with HPR-conjuncted secondary antibodies at 25 °C for 1 h. Signal detection was carried out using an enhanced chemiluminescence detection system (Pierce, Rockford, IL, USA).

We firstly detected the expression of HOXD-AS1 in 40 pairs of CC tissues (cancerous and the matched normal tissues) by qRT-PCR. The results shown in Fig. 1a revealed that among all the 40 pairs of CC tissues, 26 pairs (65.0%) expressed positive HOXD-AS1 expression (ΔΔCt> 0), and 19 pairs (47.5%) showed upregulation of HOXD-AS1 (ΔΔCt≥1) in CC tissues compared with the normal tissues. According to the classification of the clinical stages of these CC patients, we found that HOXD-AS1 expression was higher in advanced CC stages (III + IV vs. I + II; Fig. 1b). To further analyze the prognostic significance of HOXD-AS1 in CC, we divided these CC tissues into two subgroups on the basis of the median expression value, and Kaplan-Meier survival analysis was conducted. As shown in Fig. 1c, patients with higher HOXD-AS1 level presented worse overall survival than those with lower HOXD-AS1 level. Furthermore, the expression exploration in CC cell lines also showed that HOXD-AS1 expression was increased in CC cell lines compared with the normal cervical epithelial cell line (Ect1/E6E7; Fig. 1d). Therefore, these results strongly suggested that HOXD-AS1 might be implicated in CC carcinogenesis, and its higher level predicted the worse outcome of the patients.

To explore whether HOXD-AS1 exerted some effects on the mobility of CC cells, we evaluated the migration and invasion capabilities of Hela cells after transfection with its siRNAs (knockdown) or coding plasmid (overexpression). As shown in Fig. 2a and b, siRNAs specifically targeting HOXD-AS1 efficiently silenced its endogenous expression, while the coding plasmid effectively overexpressed its level in Hela cells. By transwell migration and invasion assays, we found that knockdown of HOXD-AS1 significantly repressed cell migration and invasion of Hela cells, respectively (Fig. 2c). To the opposite, forced expression of HOXD-AS1 facilitated cell migration and invasion of Hela cells (Fig. 2d). These data indicated that HOXD-AS1 itself was a metastasis promoter by accelerating migration and invasion of CC cells.

It is widely recognized that some lncRNAs serve as ceRNAs for miRNAs thus to control gene expression [13]. To screen the putative miRNAs interacting with HOXD-AS1, we utilized the online bioinformatics tool (DIANA tools, LncBase Predicted v.2, http://​carolina.​imis.​athena-innovation.​gr/​diana_​tools/​web/​index.​php) and found that miR-877-3p might bind to HOXD-AS1 (Fig. 3a). qRT-PCR showed that miR-877-3p mimics repressed the expression of HOXD-AS1, while miR-877-3p inhibitor markedly increased HOXD-AS1 expression in Hela cells (Fig. 3b). Then we constructed luciferase reporters containing the wild type miR-877-3p binding site (HOXD-AS1-wt) or the mutant type one (HOXD-AS1-mt). Co-transfection of miR-877-3p mimics and HOXD-AS1-wt largely inhibited the luciferase activity compared with the control group. However, miR-877-3p mimics failed to inhibit the luciferase activity of the HOXD-AS1-mt group (Fig. 3c). These results demonstrated that HOXD-AS1 interacted with miR-877-3p in CC cells.

To further understand the mechanism of HOXD-AS1 in CC cells, we predicted the mRNA targets of miR-877-3p by using the online Targetscan tool (http://​www.​targetscan.​org/​vert_​70/​) and found that fibroblast growth factor 2 (FGF2), a member of the fibroblast growth factor (FGF) family, might be a target of miR-877-3p (Fig. 4a). The similar luciferase reporter assay was performed, and the results showed that miR-877-3p mimics inhibited the wild type FGF2–3′-UTR (pmirGLO-FGF2–3′-UTR-wt) activity, whereas it failed to inhibit luciferase activity of the mutant type FGF2–3′-UTR (pmirGLO- FGF2–3′-UTR-mt) in which the binding site of miR-877-3p was mutated (Fig. 4a and b). qRT-PCR and western blot confirmed the targeting of FGF2 by miR-877-3p, as miR-877-3p mimics transfection led to repression of FGF2 expression in Hela cells (Fig. 4c and d). Importantly, the results also revealed that knockdown of HOXD-AS1 inhibited FGF2 expression, and HOXD-AS1 overexpression caused FGF2 upregulation at both mRNA and protein levels (Fig. 4e and f). Collectively, these results clearly implied that HOXD-AS1 could induce FGF2 expression by competitively sponging miR-877-3p in CC cells. There exists a HOXD-AS1/miR-877-3p/FGF2 ceRNA axis in CC cells.

As we had verified that HOXD-AS1 was a metastasis promoter by accelerating migration and invasion of CC cells, we then sought to determine whether this tumor promoting activity of HOXD-AS1 was mediated by the miR-877-3p/FGF2 axis. We restored the expression of miR-877-3p in HOXD-AS1-overexpressed Hela cells, and found that miR-877-3p restoration could reverse the promotion of HOXD-AS1 on CC cell migration and invasion (Fig. 5a and b). We also noticed that miR-877-3p restoration obviously downregulated FGF2 expression in HOXD-AS1-overexpressed Hela cells (Fig. 5c). To the other end, siRNA targeting FGF2 in HOXD-AS1-overexpressed Hela cells also recovered the migration and invasion capacities (Fig. 5d and e). Therefore, our findings suggested that the tumor promoting effects of HOXD-AS1 on CC cell migration and invasion were mediated by the miR-877-3p/FGF2 axis.